Aysel puts 1000 manat in a bank that adds 10% to the balance every year. In the first year she earns 100 manat, in the second already 110 manat, because the interest is charged on the new balance. Each year the balance is multiplied by the same number, 1.1, so after n years it is 1000 · 1.1ⁿ manat. After 10 years that is 2593.74 manat, while a plain “+100 manat a year” would give only 2000. Dependences in which the variable sits in the exponent are described by the exponential function.
In the lessons “Powers with integer exponents” and “Square roots, nth roots and rational exponents” we studied aⁿ for integer and rational exponents. Now the variable moves into the exponent: first we study the properties of this function, then the methods for solving exponential equations and inequalities. The next lesson, “Logarithms, their properties and the logarithmic function”, introduces a new operation that solves equations such as 2ˣ = 5.
Interactive
The exponential function and its graph
Exponential functionDefinition
A function of the form y = aˣ, where the base a is a constant (a > 0, a ≠ 1) and the variable x is in the exponent. Examples: y = 2ˣ, y = (1/3)ˣ, y = 10ˣ, y = eˣ. By contrast, y = x² is a power function, not an exponential one: there the variable is in the base.
These conditions are not arbitrary. With a < 0, values such as (−4)^(1/2) = √(−4) would make no sense, and with a = 1 we only get the constant 1ˣ = 1. For a > 0, aˣ makes sense for every real x: rational exponents are defined through roots, irrational ones by approximation — 2^√2 ≈ 2.665 is the number that the powers 2^1.4, 2^1.41, 2^1.414, … approach.
What does aˣ mean? For a natural x it is a multiplied by itself x times: 2³ = 2 · 2 · 2 = 8. Other exponents continue this pattern: each time x drops by 1, the result becomes a times smaller, so 2⁰ = 1, 2⁻¹ = 1/2, 2⁻² = 1/4. A fractional exponent is a root: 4^(1/2) = √4 = 2. The base has two bans: 1 cannot be a base, because every power of 1 is 1 and the graph would be a flat line; a negative base is not allowed either, because (−4)^(1/2) = √(−4) is not a real number.
Where
- a
- the base, a fixed positive number other than 1
- x
- the exponent, any real number
- y
- the value of the function, always positive
The exponential function: domain — all real numbers, range — (0, +∞). For a > 1 it is increasing, for 0 < a < 1 decreasing.
| x | y = 2ˣ | y = (1/2)ˣ |
|---|---|---|
| −3 | 1/8 | 8 |
| −2 | 1/4 | 4 |
| −1 | 1/2 | 2 |
| 0 | 1 | 1 |
| 1 | 2 | 1/2 |
| 2 | 4 | 1/4 |
| 3 | 8 | 1/8 |
- Domain: all real numbers; range: (0, +∞), that is, aˣ > 0. So 2ˣ = −4 and 3ˣ = 0 have no solutions.
- The graph passes through (0, 1) and (1, a), because a⁰ = 1 and a¹ = a.
- For a > 1 the function is increasing, for 0 < a < 1 it is decreasing.
- The x-axis (y = 0) is a horizontal asymptote: the graph comes as close to it as we like but never touches it.
- The graphs of y = aˣ and y = (1/a)ˣ are symmetric about the y-axis, because (1/a)ˣ = a⁻ˣ.
Interactive
Monotonicity lets us compare powers without computing them: for a > 1 the larger exponent gives the larger power, for 0 < a < 1 it is the other way round. If the bases differ, first rewrite them with the same base.
Comparing powers and finding a range
Problem
Compare:
1) 2^0.7 and 2^0.5
2) 0.3⁻² and 0.3⁻¹
3) 4^0.3 and 8^0.25
4) (1/3)^√2 and (1/3)^1.5
5) Find the range of y = 2ˣ + 3.
Show solutionHide solution
Solution
2) The base 0.3 < 1, the function is decreasing; −2 < −1 ⇒ 0.3⁻² > 0.3⁻¹ (check: 11.1 > 3.3).
3) Use the same base: 4^0.3 = 2^0.6, 8^0.25 = 2^0.75. 0.6 < 0.75 ⇒ 4^0.3 < 8^0.25.
4) √2 ≈ 1.414 < 1.5 and the base 1/3 < 1 ⇒ (1/3)^√2 > (1/3)^1.5.
5) 2ˣ takes every positive value, so 2ˣ + 3 > 3. Range: (3, +∞).
Growth and decay models. The number e
Aysel’s deposit follows this law: over equal time intervals the quantity is multiplied by the same number. If the multiplier is greater than 1, this is exponential growth (bacteria multiplying, rising prices); if it is less than 1, it is exponential decay (a car losing value, radioactive decay). Unlike linear growth, where the same amount is added each time, here the quantity grows by the same factor.
Exponential growth sooner or later overtakes any power of x, however large its exponent. Compare 2ˣ with x¹⁰: at x = 10, x¹⁰ is ten billion and 2ˣ only 1024, but from about x ≈ 58.8 on, 2ˣ is larger, and the gap then grows without limit. The reason: when x increases by 1, 2ˣ is always doubled, while x¹⁰ is multiplied by ((x + 1)/x)¹⁰, a factor that gets closer and closer to 1. This is why in computer science an algorithm with 2ⁿ steps is hopeless for large n, while one with n¹⁰ steps is still called “polynomial”.
| x | 2ˣ | x¹⁰ |
|---|---|---|
| 10 | 1024 | 10¹⁰ |
| 30 | ≈ 1.07 · 10⁹ | ≈ 5.90 · 10¹⁴ |
| 60 | ≈ 1.15 · 10¹⁸ | ≈ 6.05 · 10¹⁷ |
| 100 | ≈ 1.27 · 10³⁰ | 10²⁰ |
Where
- N₀
- the initial amount (at t = 0)
- a
- the multiplier per period: a = 1 + p/100 for growth by p%, a = 1 − p/100 for a decrease by p%, a = 2 for doubling, a = 1/2 for halving
- T
- the length of one period (year, hour, day…)
- t
- the time elapsed; t/T is the number of periods
Every time T passes, the quantity is multiplied by a. Compound interest: S = S₀ · (1 + p/100)ⁿ (T = 1 year, n = number of years). Radioactive decay: m = m₀ · (1/2)^(t/T), where T is the half-life.
Count the periods
Problem
Bacteria double every 3 hours. There are 100 of them now. How many will there be after 9 hours?
Show solutionHide solution
Solution
In 9 hours there are 9 ÷ 3 = 3 periods, and in each one the number is multiplied by 2: 100 · 2 · 2 · 2 = 100 · 2³ = 800.
This is exactly the formula: N₀ = 100, a = 2, t/T = 9/3 = 3.
Interest, depreciation, multiplication and decay
Problem
1) Aysel deposits 1000 manat at 10% compound interest per year for 3 years. How much will she have at the end?
2) A car worth 20 000 manat loses 20% of its value every year. What is it worth after 3 years?
3) Bacteria double every 20 minutes. How many bacteria come from 500 bacteria after 2 hours?
4) The half-life of iodine-131 is about 8 days. How much of 80 mg of iodine is left after 24 days?
Show solutionHide solution
Solution
2) a = 1 − 0.2 = 0.8: 20 000 · 0.8³ = 20 000 · 0.512 = 10 240 manat.
3) T = 20 min, t = 120 min ⇒ t/T = 6: N = 500 · 2⁶ = 500 · 64 = 32 000.
4) t/T = 24 ÷ 8 = 3: m = 80 · (1/2)³ = 80 ÷ 8 = 10 mg.
What happens if interest is added more often than once a year? Suppose 1 manat is deposited at 100% a year and the interest is added n times a year, each time at the rate 100%/n. At the end of the year there are (1 + 1/n)ⁿ manat. As n grows, this number grows too, but not without bound: it approaches a certain number.
Simply put, the bank pays interest on interest. If 100% is added once, 1 manat becomes 2 manat. If 50% is added twice, we get 1.5 · 1.5 = 2.25 manat. The more often (and in smaller slices) the interest is added, the more money there is, but the extra gain keeps shrinking: monthly interest gives about 2.61, daily about 2.71. However often the interest is added, the result never goes past one particular number, about 2.718. This number is called e.
| Interest added | n | (1 + 1/n)ⁿ |
|---|---|---|
| once a year | 1 | 2 |
| every half year | 2 | 2.25 |
| monthly | 12 | 2.613035… |
| daily | 365 | 2.714567… |
| a million times | 1 000 000 | 2.718280… |
Where
- e
- Euler’s number, an irrational number and the base of the natural logarithm
- n
- how many times a year the interest is added
The function y = eˣ with base e is often called simply the exponential function. When a quantity changes at every moment at a rate proportional to its size (bacteria, populations, radioactive matter), it is expressed through powers of e.
The sum of this series is e, and it converges very fast: the first 11 terms already give 2.7182818 (7 correct decimal places), while (1 + 1/n)ⁿ gives only 5 correct decimal places even for n = 1 000 000. Here k! = 1 · 2 · … · k.
Why does (1 + 1/n)ⁿ have a limit at all? By the binomial theorem, (1 + 1/n)ⁿ = 1 + 1 + (1 − 1/n)/2! + (1 − 1/n)(1 − 2/n)/3! + … . As n grows, every term grows and new terms appear, so the sequence is increasing. Each term is at most 1/k!, which is at most 1/2ᵏ⁻¹, so the sum is less than 1 + 1 + 1/2 + 1/4 + … = 3. An increasing sequence that is bounded above has a limit — and this limit is e. In calculus, eˣ is the only function (up to a constant factor) equal to its own derivative: (eˣ)′ = eˣ. That is why e describes every process whose rate of growth is proportional to the amount.
Where
- N₀
- the initial amount
- k
- the rate constant: k > 0 for growth, k < 0 for decay
- t
- time
The model of continuous growth and decay. Powers of e are found with a calculator: e ≈ 2.718, e⁻¹ ≈ 0.368, e^0.1 ≈ 1.105.
Models with base e
Problem
1) A lab has 2000 bacteria and N(t) = 2000 · e^(0.5t) (t in hours). How many bacteria are there after 2 hours?
2) The mass of a substance decreases as m(t) = 100 · e^(−0.1t) grams (t in days). How much is left after 10 days?
3) 1000 manat is kept for 1 year at 10% continuous interest: S = 1000 · e^0.1. Compare with interest added once a year.
Show solutionHide solution
Solution
2) m(10) = 100 · e⁻¹ ≈ 100 · 0.368 ≈ 36.8 g.
3) S = 1000 · e^0.1 ≈ 1105.17 manat, while once a year gives 1000 · 1.1 = 1100 manat. Continuous interest adds only 5.17 manat more: the growth is not unlimited, e caps it.
Every exponential model can be rewritten with base e, because a = e^(ln a) (the natural logarithm ln is the topic of the next lesson): aᵗ = e^(t · ln a). For decay with half-life T this gives m = m₀ · e^(−kt) with k = ln 2 / T ≈ 0.693/T. For growth by p% a year, the doubling time is n = ln 2 / ln(1 + p/100); for small p this is about 70/p years — bankers use the “rule of 72”: 72/p.
Radiocarbon and the rule of 72
Problem
1) Radiocarbon (C-14) has a half-life of about 5730 years. A bone contains 25% of its original C-14. How old is it?
2) How long does it take a deposit to double at 7% a year?
Show solutionHide solution
Solution
2) We need 1.07ⁿ ≥ 2: 1.07¹⁰ ≈ 1.967 is not yet enough, 1.07¹¹ ≈ 2.105 is. The exact time is ln 2 / ln 1.07 ≈ 10.24 years; the rule of 72 gives 72 ÷ 7 ≈ 10.3 — very close.
Exponential equations: four main methods
Exponential equationDefinition
An equation with the unknown in the exponent: 2^(x + 1) = 32, 9ˣ − 4 · 3ˣ + 3 = 0. Every method has the same goal: to bring the equation to the form a^f(x) = a^g(x).
Where
- a
- the same base on both sides
- f(x), g(x)
- the expressions in the exponents
Why it holds: aˣ is monotonic and takes each value only once, so equal powers have equal exponents. The equation aˣ = b has exactly one solution for b > 0 and none for b ≤ 0.
Look at the bases
Try to write all powers with one base: 4 = 2², 0.25 = 2⁻², √2 = 2^(1/2), 1/27 = 3⁻³.
Several terms with one base
Factor out the power with the smallest exponent: 3^(x + 2) − 3ˣ = 3ˣ(9 − 1).
a^(2x) together with aˣ
Substitute t = aˣ, write down t > 0 and solve the quadratic equation.
Two different bases: 4ˣ, 6ˣ, 9ˣ
It is a homogeneous equation: divide by 9ˣ and put t = (2/3)ˣ.
Check
Substitute the root into the original equation, or at least check that t > 0.
| What do you see? | Method | Tiny example |
|---|---|---|
| both sides are powers of one number | common base | 2ˣ = 32 = 2⁵ ⇒ x = 5 |
| one base, the exponents differ by a number | factoring out | 3^(x + 1) + 3ˣ = 36 ⇒ 4 · 3ˣ = 36 ⇒ x = 2 |
| a^(2x) together with aˣ | t = aˣ > 0 | 4ˣ − 2ˣ − 2 = 0 ⇒ t² − t − 2 = 0 ⇒ t = 2 ⇒ x = 1 |
| 4ˣ, 6ˣ, 9ˣ — two bases | divide by 9ˣ, t = (2/3)ˣ | 4ˣ = 9ˣ ⇒ (4/9)ˣ = 1 ⇒ x = 0 |
Method 1: a common base
Problem
Solve:
1) 2^(x + 1) = 32
2) 9ˣ = 27
3) (1/5)ˣ = 125
4) 4^(x² − x) = 16
5) 2ˣ = 4√2
Show solutionHide solution
Solution
2) 9ˣ = 3^(2x), 27 = 3³ ⇒ 2x = 3 ⇒ x = 3/2.
3) 1/5 = 5⁻¹ ⇒ 5^(−x) = 5³ ⇒ x = −3.
4) 16 = 4² ⇒ x² − x = 2 ⇒ x² − x − 2 = 0 ⇒ x = 2 or x = −1.
5) 4√2 = 2² · 2^(1/2) = 2^(5/2) ⇒ x = 5/2.
Where
- k
- a constant (the difference of the exponents)
Powers whose exponents differ by a constant share the factor aˣ. Factor it out and an ordinary number is left in the brackets.
Method 2: factoring out
Problem
Solve:
1) 3^(x + 2) − 3ˣ = 72
2) 2^(x + 1) + 2^(x − 1) = 20
3) 5ˣ + 5^(x + 1) = 150
Show solutionHide solution
Solution
2) The smallest exponent is x − 1: 2^(x − 1)(2² + 1) = 20 ⇒ 2^(x − 1) = 4 ⇒ x − 1 = 2 ⇒ x = 3.
Check: 2⁴ + 2² = 16 + 4 = 20.
3) 5ˣ(1 + 5) = 150 ⇒ 5ˣ = 25 ⇒ x = 2.
Where
- t
- the new variable; always positive, because aˣ > 0
An equation made of a^(2x), aˣ and a constant becomes a quadratic equation in t. Since a^(−x) = 1/t, equations with aˣ + a^(−x) are solved the same way.
Method 3: substitution
Problem
Solve:
1) 9ˣ − 4 · 3ˣ + 3 = 0
2) 4ˣ − 2ˣ − 12 = 0
3) 2ˣ + 2^(2 − x) = 5
Show solutionHide solution
Solution
3ˣ = 1 ⇒ x = 0; 3ˣ = 3 ⇒ x = 1. Answer: 0 and 1.
2) t = 2ˣ > 0: t² − t − 12 = 0 ⇒ t = 4 or t = −3.
Reject t = −3, because 2ˣ cannot be negative. 2ˣ = 4 ⇒ x = 2.
3) 2^(2 − x) = 4/2ˣ. t = 2ˣ > 0: t + 4/t = 5 ⇒ t² − 5t + 4 = 0 ⇒ t = 1 or t = 4.
Answer: x = 0 and x = 2.
Where
- a, b
- two different positive bases (for example 2 and 3)
- A, B, C
- numerical coefficients
A homogeneous equation: every term has the same “degree” (a^(2x), aˣbˣ, b^(2x)). Since b^(2x) is never 0, we may divide by it and get a quadratic equation in one unknown. The simplest case: aˣ = bˣ ⇔ (a/b)ˣ = 1 ⇔ x = 0.
Method 4: homogeneous equations
Problem
Solve:
1) 5ˣ = 7ˣ
2) 3 · 4ˣ − 5 · 6ˣ + 2 · 9ˣ = 0
3) 9ˣ + 6ˣ = 2 · 4ˣ
Show solutionHide solution
Solution
2) 4ˣ = (2ˣ)², 9ˣ = (3ˣ)², 6ˣ = 2ˣ · 3ˣ. Divide by 9ˣ: 3 · (2/3)^(2x) − 5 · (2/3)ˣ + 2 = 0.
t = (2/3)ˣ > 0: 3t² − 5t + 2 = 0 ⇒ t = 1 or t = 2/3.
(2/3)ˣ = 1 ⇒ x = 0; (2/3)ˣ = 2/3 ⇒ x = 1. Answer: 0 and 1.
Check (x = 1): 12 − 30 + 18 = 0.
3) Divide by 4ˣ: (3/2)^(2x) + (3/2)ˣ − 2 = 0. t = (3/2)ˣ > 0: t² + t − 2 = 0 ⇒ t = 1 or t = −2 (rejected).
(3/2)ˣ = 1 ⇒ x = 0.
Monotonicity instead of algebra
Problem
Solve 3ˣ + 4ˣ = 5ˣ.
Show solutionHide solution
Solution
Answer: x = 2 — the only root. The homogeneous-equation method does not work here: there are three different bases.
Reciprocal powers
Problem
Solve (√(2 + √3))ˣ + (√(2 − √3))ˣ = 4.
Show solutionHide solution
Solution
t + 1/t = 4 ⇒ t² − 4t + 1 = 0 ⇒ t = 2 ± √3.
2 + √3 = (√(2 + √3))² ⇒ x = 2; 2 − √3 = 1/(2 + √3) = (√(2 + √3))⁻² ⇒ x = −2.
Answer: x = ±2.
Exponential inequalities
In an inequality we also bring both sides to one base, but when we move to the exponents one question decides everything: is the function increasing or decreasing? An increasing function gives a larger value for a larger argument, so the sign stays; for a decreasing function the sign flips.
Think of the graph as a staircase. 2ˣ is a staircase going up: the further right, the higher. So 2ˣ > 2³ simply means “x is to the right of 3”: x > 3. (1/2)ˣ is a staircase going down: the further right, the lower. So (1/2)ˣ > (1/2)³ means “x is to the left of 3”: x < 3. Check with x = 2: (1/2)² = 1/4 is indeed greater than (1/2)³ = 1/8.
Where
- a
- the common base
- f(x), g(x)
- the expressions in the exponents
If the base is less than 1, the sign flips. If the right-hand side is negative or zero, nothing needs solving: aˣ > −3 holds for every x, and aˣ < 0 never holds.
Inequalities with a common base
Problem
Solve:
1) 2^(x − 1) > 8
2) (1/3)ˣ ≥ 27
3) 0.5^(x² − 2x) ≥ 1/8
4) 3^(x²) < 9^(x + 4)
Show solutionHide solution
Solution
2) 27 = (1/3)⁻³, the base 1/3 < 1, the sign flips: x ≤ −3.
3) 1/8 = 0.5³, the base 0.5 < 1: x² − 2x ≤ 3 ⇒ (x − 3)(x + 1) ≤ 0 ⇒ x ∈ [−1, 3].
4) 9^(x + 4) = 3^(2x + 8), the base 3 > 1: x² < 2x + 8 ⇒ (x − 4)(x + 2) < 0 ⇒ x ∈ (−2, 4).
The last two were finished with the interval method from the lesson “Quadratic inequalities and the interval method”.
Inequalities by substitution
Problem
Solve:
1) 4ˣ − 5 · 2ˣ + 4 < 0
2) 9ˣ − 2 · 3ˣ − 3 > 0
Show solutionHide solution
Solution
2⁰ < 2ˣ < 2² and the base 2 > 1 ⇒ 0 < x < 2.
2) t = 3ˣ > 0: (t − 3)(t + 1) > 0. Since t > 0, t + 1 > 0, so t − 3 > 0, that is, t > 3.
3ˣ > 3¹ ⇒ x > 1.
The rationalization method. For a > 1 both brackets have the same sign; for 0 < a < 1 the sign flips — the rule of this lesson written as one inequality. For a = 1 both sides are false. The method shows its strength when the base itself depends on x: then one interval-method inequality replaces several cases (a > 1, a = 1, 0 < a < 1).
A variable base
Problem
Solve (x² + x + 1)ˣ < 1.
Show solutionHide solution
Solution
For x ≠ 0, x² > 0, so x + 1 < 0: x < −1.
Check: x = −2 — the base is 3 and 3⁻² = 1/9 < 1 ✓; x = −0.5 — the base is 0.75 and 0.75^(−0.5) ≈ 1.155 > 1, not a solution ✓.
Remember
- y = aˣ (a > 0, a ≠ 1): domain — all real numbers, range — (0, +∞); the graph passes through (0, 1).
- For a > 1, aˣ is increasing, for 0 < a < 1 decreasing; the graphs of y = aˣ and y = (1/a)ˣ are symmetric about the y-axis.
- Growth and decay model: N(t) = N₀ · a^(t/T); e ≈ 2.718 is the number that (1 + 1/n)ⁿ approaches; the continuous model is N₀ · e^(kt).
- a^f(x) = a^g(x) ⇔ f(x) = g(x). Methods: a common base, factoring out, the substitution t = aˣ > 0, dividing a homogeneous equation by b^(2x).
- In an inequality the sign stays for a > 1 and flips for 0 < a < 1; remember that aˣ > 0.